使用tf.add_to_collection和tf.get_collection获取中间量

tf.add_to_collection–向当前计算图中添加张量集合

tf.get_collection–返回当前计算图中手动添加的张量集合

#!/usr/bin/python
# coding:utf-8

import tensorflow as tf
v1 = tf.get_variable('v1', shape=[3], initializer=tf.ones_initializer())
v2 = tf.get_variable('v2', shape=[5], initializer=tf.random_uniform_initializer(maxval=-1., minval=1., seed=0))
# 向当前计算图中添加张量集合
tf.add_to_collection('v', v1)
tf.add_to_collection('v', v2)

with tf.Session() as sess:
    sess.run(tf.global_variables_initializer())
    # 返回当前计算图中手动添加的张量集合
    v = tf.get_collection('v')
    print v
    print v[0].eval()
    print v[1].eval()

输出：

[<tf.Variable 'v1:0' shape=(3,) dtype=float32_ref>, <tf.Variable 'v2:0' shape=(5,) dtype=float32_ref>]
[ 1.  1.  1.]
[ 0.79827476 -0.9403336  -0.69752836  0.90343738  0.90295386]

 

原文：https://blog.csdn.net/akadiao/article/details/78521760